WO1994008598A1 - Liver reserve cells - Google Patents

Abstract

The present invention relates to liver reserve or stem cells. In particular, it relates to the isolation, characterization, culturing, and uses of liver reserve cells. Liver reserve cells isolated by density gradient centrifugation can be distinguished from other liver parenchymal cells by their morphology, staining characteristics, high proliferative activity and ability to differentiate in vitro. In long-term cultures described herein, these cells expand in numbers and differentiate into morphologically mature liver parenchymal cells, capable of mediating liver-specific functions. Therefore, isolated liver reserve cells may have a wide range of applications, including, but not limited to, their uses as vehicles of exogenous genes in gene therapy, and/or to replace and reconstitute a destroyed, infected, or genetically deficient mammalian liver by transplantation.

Description

LIVER RESERVE CELLS

1. INTRODUCTION The present invention relates to liver reserve or stem cells. In particular, it relates to the isolation, characterization, culturing, and uses of liver reserve cells. Liver reserve cells isolated by density gradient centrifugation can be distinguished from other liver parenchymal cells by their morphology, staining characteristics, high proliferative activity and ability to differentiate in vitro. In long-term cultures described herein, these cells expand in numbers and differentiate into morphologically mature liver parenchymal cells, capable of mediating liver-specific functions. Therefore, isolated liver reserve cells may have a wide range of applications, including, but not limited to, their use as vehicles of exogenous genes in gene therapy, and/or to replace and reconstitute a destroyed, infected, or genetically deficient mammalian liver by transplantation.

2. BACKGROUND OF THE INVENTION The liver is a dynamic organ that plays an important role in a variety of physiological processes. The complex functions of the liver include metabolism, storage, excretion, secretion of plasma proteins such as albumin and detoxification of harmful substances by enzymes of the cytochrome P-450 system. In addition, the usually quiescent liver is also capable of remarkable mitotic activities under certain circumstances. 2.1. LIVER CELLS The major cell population of the liver is the parenchymal cells (PC) , also known as hepatocytes. ₅ The liver also contains several other cell types εuch as endothelial cells, adipocytes, fibroblastic cells and Kupffer cells, collectively referred to as stromal (littoral) cells. The ability of liver cells to undergo rapid regeneration, when the liver is damaged

₁₀ or partially removed, has led to speculation for the existence of a population of stem cells or reserve cells, capable of self-renewal and differentiation. However, prior to the present invention, such liver reserve cells had never been identified or their _l5 characteristics described.

"Oval" cells have been described in the adult mammalian liver. These cells have a high nuclear to cytoplasmic size ratio, are approximately 40% of the diameter of freshly isolated PC, express ₀ enzyme activities consistent with those of fetal hepatocytes, and have a relatively high proliferation rate (Tsao et al., 1984, Exp. Cell Res. 154:38-52; Farber et al., 1956, Cancer Res. 16:142-149). However, animals usually must be treated in vivo with

25 ethionine, 2-acetylaminofluorene, or other carcinogens to generate these cells. Some investigators have hypothesized that oval cells may be a manifestation of PC retrodifferentiation (Grisham, 1980, Ann. N.Y. Acad. Sci. 349:128-137; Firminger, 1955, J. Nat.

₃₀ Cancer Instit. 15:1427-1441) but they may resemble endothelial cells more closely (Fausto et al., 1987, In: Cell Separation: Methods and selected applications, Vol. 4, T.G. Pretlow II and T.P. Pretlow, editors, Academic Press, London, pp 45-78) . ₅ The role of oval cells as putative stem cells or reserve cells has never been established, and it has been the subject of a number of investigations (Fausto et al. , 1987, In: Cell Separation: Methods and selected applications. Vol. 4, T.G. Pretlow II and ₅ T.P. Pretlow, editors. Academic Press, London, pp 45-78) .

2.2. LIVER CULTURES In an attempt to study the diverse liver _l0 functions and the cell types responsible therefor, in vitro cultures of liver cells have been prepared from humans as well as from experimental animals. Primary cultures of rat hepatocytes have been used extensively to study the effects of potential toxins on enzyme

15 leakage, metabolism, and cellular membranes (Grisham, 1979, Int. Rev. Exp. Pathol. 20:123-210; Acosta and Mitchell, 1981, Biochem. Pharmacol. 30:3225-3230). However, such culture systems have a number of drawbacks, and none have provided for the ₀ proliferation of liver PC or the identification of liver reserve cells.

In vitro, adult hepatocytes proliferate for only short time periods, although their ability to produce albumin and display cytochrome P-450 enzyme 5 activity may be prolonged if they are co-cultured with other liver-derived extracellular matrix substances or with certain combinations thereof. In liquid culture, the viability of hepatocytes and the ability of these cells to manifest inducible cytochrome P-450 enzyme ₀ activity decline as a function of time (Sirica and Pitot, 1980, Pharmacol. Rev. 31:205-228). In addition, cell division usually is limited to the first 24-48 hr of culture (Clayton and Darnell, 1983, Mol. Cell Biol. 3:1552-1561; Chapman et al., 1973, J.

₃5 Cell Biol. 59:735-747). The viability of adherent hepatocytes in monolayer cultures persists for - A -

somewhat longer periods but specialized activity is also lost rapidly (Deschenes et al., 1980, In Vitro 16:722-730) . _s Towards the goal of enhancing hepatocyte growth and prolonging liver-specific functions in vitro, hepatic cells have been cultured on various matrices including type I collagen plates and membranes (Michalopoulos and Pitot, 1975, Exp. Cell ₀ Res. 94:70-78), homogenized liver biomatrix (Reid et al., 1980, Ann. N.Y. Acad. Sci. 349:70-76), in collagen type IV or laminin-rich gels (Biεsell et al., 1987, J. Clin. Invest. 79:801-812), sandwiched between two layers of type I collagen (Dunn et al., 1989, FASEB J. 3:174-177), and on plates coated with fibronectin or the other extracellular matrix proteins (Deschenes et al., 1980, In Vitro 16:722-730). All of these methods have been reported to extend the functional life of hepatocytes in vitro to some o extent.

Substantial improvements in this regard were produced by culturing PC with various types of non-parenchyma1 stromal or littoral hepatic cells or non-hepatic stromal cells. Both human and rat 5 hepatocytes which were co-cultured with liver endothelial cells of the same species maintained specific functions for weeks in culture, although they did not undergo a significant expansion in numbers (Guguen-Guilluozo, et al., 1983, Exp. Cell Res. ₀ 143:47-54; Begue et al. , 1983, Bioche . Pharmacol.

32:1643-1646). Rat hepatocytes which were co-cultured with human fibroblasts (Kuri-Harcuch and Mendoza-Figueroa, 1989, Differentiation 41:148-157) and endothelial cells (Begue et al., 1983, Biochem. ₅ Pharmacol. 32:1643-1646) were reported to sustain cytochrome P-450 activity for more than 10 days. Thus, these mixed hepatocyte co-culture systems may provide microenvironments similar to those in vivo by optimizing cell-cell interactions. In addition, various PC functions may be regulated and/or optimized by other hepatic cells. For example, Kupffer cell secretory products have been reported to modulate PC cytochrome P-450 enzyme activity (Peterson and Renton, 1984, J. Pharmacol. Exp. Ther. 229:299-304). The attachment of PC to fibroblasts is evidently contingent upon the secretion of specialized extracellular matrix substances by Kupffer cells (Michalopoulos et al., 1979, In Vitro 15:769-806). Hepatic endothelial cells also may produce important components of the extracellular matrix

(Guguen-Guilluozo, et al., 1983, Exp. Cell Res. 143:47-54), and adipocytes may provide the requisite raw materials for the renewal of cell membranes in metabolically-active hepatocytes. Although the viability and functional activities of cultured hepatic PC can be prolonged in vitro if the cells are co-cultured with non-parenchyma1 liver stromal cells, support cells from other tissues, or their secretory products, PC proliferation is limited or absent in these systems. Mitoses in co-cultures of hepatic cells have been ascribed primarily to non-parenchyma1 elements (Guguen-Guilluozo, et al., 1983, Exp. Cell Res. 143:47-54). Several reports indicate that non-parenchyma1 liver cells may express functions similar to hepatocytes (Grisham, 1980, Ann. N.Y. Acad. Sci. 349:128-137) although the nature of theεe non-PC has not been unequivocally established.

The growth of rat hepatocytes has been particularly enhanced when cultured on a three-dimensional template consisting of hepatic-derived stromal cells attached to a nylon filtration screen (Naughton and Naughton, 1991, United States Patent No. 5,032,508). The stromal compartment contains all of the adherent cells found in liver tissues including Kupffer cells, vascular and bile duct endothelial cells, fibroblaεtε, and some adipocyte-like cells. These cells elaborate a matrix similar in some respects to that observed in liver in vivo and support long-term growth of PC and their liver specific functions in vitro. However, prior to the present invention, the existence of liver reserve cells in εuch cultures had never been established.

3. SUMMARY OF THE INVENTION

The present invention relates to liver reserve cells, a method of isolating and culturing liver reserve cells, and a method of using the liver reserve cells with or without exogenous genetic materialε in tranεplantation or implantation into an individual with a specific liver disorder.

The invention is based, in part, on Applicants' discovery that rat liver PC can be grown in a culture system to suεtain liver-εpecific functions for over 60 days, and the culture contains all of the cell types found in the liver in vivo. The co-cultures composed of liver PC grown upon liver stromal cells which have attached to a nylon screen, when placed in liquid medium, become εuεpended between the bottom of the flask and the surface of the medium, enhancing the three-dimensional growth effect. Hepatocellular DNA synthesis, which persiεts for >7 weeks in vitro, is enhanced by supplementing the medium with εaturated transferrin and/or conditioning the medium with sera derived from the hepatic veins of partially hepatectomized rats. Liver specific functions such as albumin synthesis and cytochrome P- 450 enzyme activity are evident for as long as 60 days in culture. In addition, the expression of both class I MHC antigens on cultured hepatic parenchyma and MHC class II antigens on Kupffer cells declined as a function of time in vitro. Sections through pellets of adherent zone cells revealed normal parenchymal cell architecture. ^{In the} course of iεolating and culturing liver PC, a population of previously unknown large acidophilic hepatic cells, which have a higher proliferation rate than other hepatic cells, were isolated using density gradient centrifugation. Theεe cultured acidophilic hepatocytes undergo cell division in the culture syεtem described above, and cell cycle analysis suggests that they are PC rather than stromal cells because of their total DNA content. Medium containing transferrin εaturated with ferric iron and supplemented with ferrouε sulphate and ferric citrate and/or conditioned with sera from hepatectomized rats enhance the mitotic indices of inocula of either acidophilic cells alone or mixed acidophilic and mature PC on suspended nylon screen cultures. Most importantly, these acidophilic cells are able to develop into hepatocytes in culture and perform liver-specific functions, and thus, are believed to be liver reserve cells.

The invention is described by way of examples in which rat liver reserve cells are isolated and their cytologic and biologic properties characterized. A relatively homogeneous (>90%) population of liver reserve cells can be maintained in cultures, and shown to retain proliferative and differentiative capabilities. A wide variety of uses for the liver reserve cells are encompassed by the invention described herein. In particular, the high proliferative rate of these cells allows them to be ideal recipientε for εtable inεertion of exogenouε geneε in gene therapy of various liver disorderε.

4. BRIEF DESCRIPTION OF THE DRAWINGS Figure 1. Photomicrograph of a cytoεmear of the pellet of a 70% Percoll denεity centrifugation. Mature parenchymal cells stain darkly; the stain had to be applied for a longer time in order to viεualize the larger, lightly staining acidophilic cells.

Figure 2. Photomicrograph of a cytosmear of acidophilic cells from the discontinuous "neat Percoll" procedure. Diff-Quik stain. Original magnification - 1000X. These cells are 25-30% larger than other cells in the preparation, are vacuolated and faintly staining, display numerous, large nύcleoli, and have a lower nuclear : cytoplasmic size ratio than other hepatic cells of the isolate.

Figure 3. Cytosmear of morphologically mature parenchymal cells in the non-adherent zone of monolayer cultureε initiated with acidophilic cells (10 day culture) .

Figure 4. Mature parenchymal cells adhering to and posεibly emanating from an adherent zone of acidophilic cells in monolayer culture.

Figure 5. A. Inverted phase micrograph of a liver co-culture established by inoculating cells from a 70% Percoll density gradient onto a sub-confluent layer of liver stromal cells on nylon εcreen.

B. Inverted phase micrograph of a liver co-culture establiεhed by inoculating acidophilic liver reεerve cells isolated using a discontinuous Percoll density gradient procedure.

Figure 6. Cell-cycle analysis of liver co-cultures under different conditions as measured in isolated nuclei stained with propidium iodide by flow cytometry.

Figure 7. Tritiated thy idine incorporation into liver cells under various culture conditions.

Figure 8. Transmission electron micrograph of a section through a pellet of en^yme-diεεociated adherent zone cells of a nylon εcreen liver cell culture 85 dayε after inoculation. Patent mitochondrial ultraεtructure and metabolically active rough endoplasmic reticulum membranes are seen. Original magnification = 2,800X.

Figure 9. A. Mean albumin secretion (+ 1 standard error of the mean) into the medium by cells from the adherent zones of suspended nylon screen liver cultures of various ages.

B. Immunoperoxidase stain for albumin in a

40 day liver co-culture. Figure 10. Diagram of the metabolic conversion of ethoxyfluorescein ethyl eεter (EFEE) to fluoreεcein(F) by cells with cytochrome P-450 enzyme activity. Cytochrme P-450 enzyme activity in liver cultures of various ages determined .by EFEE to F conversion as meaεured by flow cytometry.

Figure 11. A. Area plot depicting relative percentages of various hepatic cells in the adherent zones of εuspended nylon screen cultures.

B. Cytosmear of adherent zone of a liver co-culture at 40 days showing the persiεtence of acidophilic cells aε an entity in εtromal cell-associated culture.

Figure 12. Hematoxylin and eosin stained section through a nylon screen co-culture.

Figure 13, cultures for toxicity measurementε.

Neutral viability assay εhowing diminiεhed viability of co-cultured hepatocyteε with increaεing concentrationε of alcohol.

B. Diminiεhed tritiated thymidine incorporation into DNA of cultured liver cells related to a dose of 5- Fluorouracil. C. Activation of cyclophosphamide and benzene by liver cells co-cultured with bone marrow cells. Under these conditions, bone marrow viability (MTT assay) diminisheε with increaεing doses of cyclophosphamide and benzene.

5. DETAILED DESCRIPTION OF THE INVENTION The present invention relates to liver reεerve cells, to methods of isolating and culturing these cells, and to methods of using the same.

Liver reserve cells are separated from other rat liver PC by differential gradient centrifugation, and they exhibit unique phenotypic characteristics as follows. The reserve cells are larger than all typical PC with a low nuclear size to cytoplaεmic size ratio, and have two to three prominent nucleoli per nucleuε. Theεe cellε primarily take up the acid component of a combination of acidic/baεic stain, while other PC are strongly baεophilic. Liver reεerve cellε have a high rate of proliferation (a doubling time of 24-28 hourε) in monolayer culture and persist for even longer periods in nylon screen co-cultures with εtromal cellε. In culture, they are rapidly and firmly adherent to plastic aε well as to other "mature" PC. Further, they are capable of differentiating into cells with typical PC morphology and function. After a rapid growth period in monolayer culture, cells apparently derive from the adherent acidophilic reserve cell culture and become detached from other cells; these have a higher nuclear to cytoplasmic εize ratio and accumulate the baεic component of the εtain. The liver-εpecific functional activities of these cells include the secretion of albumin and the expreεsion of inducible cytochrome P-450 enzymes.

The invention is discussed in more detail in ₅ the subεections below, solely for purposes of description and not by way of limitation. For clarity of discussion, the εpecific procedureε and methods described herein are exemplified using rat liver preparations, but they are merely illustrative for the ₀ practice of the invention. Analogous procedures and techniques are equally applicable to all mammalian εpecieε, including human εubjects. In fact, human liver reserve cellε have been iεolated from human liver preparations, following similar procedures and 5 shown to display a phenotype conεistent with the rat acidophilic liver reserve cells illustrated herein.

5.1. ISOLATION OF LIVER RESERVE CELLS The present invention relates to liver o reserve cells which are present in low numbers in a normal liver, occupying 2-5% of the total liver PC population. The larger size of these cellε serves as a convenient basis for their separation from other liver cells. Thus, theεe cellε may be iεolated from a 5 liver by differential denεity gradient centrifugation as described in Section 6.1.1., infra. The large cells obtained at the upper portion of the interface zone following Percoll diεcontinuous gradient centrifugation consists of greater than 90% large ₀ acidophilic liver reserve cells. This level of enrichment is generally acceptable for the various uses of these cells described in Section 5.4., infra.

Alternatively, liver reserve cells may be isolated by subjecting liver cell preparations to ₅ lectin chro atography, affinity chromatography involving positive and negative selection, repetitive density gradient centrifugation, or a combination thereof. For example, liver reserve cells obtained in Section 6.1.1., infra, may be negatively εelected by 5 panning using antibodies to remove fibroblastε, endothelial cells, Kupffer cells and adipocytes. Examples of εuch antibodieε include anti-MHC class II antigens and anti-vW factor VIII εpecific for Kupffer cellε and endothelial cellε, respectively. These antibodies may be applied in any combination repeatedly or in a sequential manner. Upon binding to the antibodies, the cells may be removed by adsorption to a solid surface or column coated with a second step antibody; i.e., an anti-mouse antibody, if the primary ₅ antibody is of mouse origin; or if the antibodies are conjugated with biotin, the antibody-bound cells can be removed by an avidin-coated column; or if the antibodies are conjugated to magnetic beads, the cells expressing antigens recognized by the antibodies can o b removed in a magnetic field. A similar procedure may be followed for positive selection, except that antibodies directed to liver reεerve cell-specific surface markers are utilized to sort out the desired cell population. 5

5.2. LONG-TERM CULTURE OF LIVER RESERVE CELLS AND OTHER PARENCHYMAL CELLS

Liver, although a mitotically quiescent organ, has the potential to regenerate after total hepatectomy or partial chemical deεtruction (Bucher, 0

1963, Int. Rev. Cytol. 15:245-300; Naughton et al.,

1977, Science 196:301-302) . That this phenomenon may be related to liver reserve cells is important in the understanding of the induction of both hepatic regeneration and hepatocarcinogenesis. In a εpecific 5 embodiment by way of example in Example 6, infra. liver PC or acidophilic liver reserve cells are shown to grow on nylon screen seeded with stromal cells and persist for extended periods of time in culture. Additionally, the preεence of εtromal cellε or the matrix proteinε secreted by these elements iε necessary for the prolongation of hepatocyte cytochrome P-450 enzyme function (Deschenes et al., 1980, In Vitro 16:722-730; Michalopoulos and Pitot, 1975, Exp. Cell Reε. 94:70-78; Reid et al., 1980, Ann. N.Y. Acad. Sci. 349:70-76; Bissell et al., 1987, J. Clin. Inveεt. 79:801-812; Dunn et al., 1989, FASEB J. 3:174-177; Guguen-Guilluozo, et al., 1983, Exp. Cell Reε. 143:47-54; Begue et al., 1983, Biochem. Pharmacol. 32:1643-1646; Kuri-Harcuch and Mendoza-Figueroa, 1989, Differentiation 41:148-157). Fibronectin, laminin, and collagen are produced by the hepatic stromal cellε onto which the PC or liver reεerve cellε can be εeeded. The levels of these substanceε, aε eεtimated by the fluoreεence intenεity after labelling with primary antibodieε and fluorochrome-conjugated εecondary antibodieε are higher than that obεerved in liver organ slices. The increased syntheεiε of these extracellular matrix proteins mimicε their expression pattern in fetal tissues.

Extracellular matrix εubstances produced by stromal cells in the liver co-cultures may influence cell division and gene expresεion in PC and liver reεerve cells (Tonomura et ai., 1987, J. Cell Physiol. 130:221-227; Sudhakaran et al., 1986, Exp. Cell Reε. 167:505-516). In this regard, DNA synthesis rates for hepatocytes in vitro are higher on fibronectin vs. laminin-rich substrates (Tonomura et al., 1987, J. Cell Physiol. 130:221-227) and the synthesis of α-fetoprotein and albumin by cultured liver cells is greatest in association with type IV collagen (Sudhakaran et al., 1986, Exp. Cell Res. 167:505-516). Similar to the liver in vivo, the extracellular matrix of hepatocyte co-cultures of the present invention contains fibronectin, laminin, and collagen IV. Production of TGFjS, fibrinogen and other liver-εpecific proteins by the cultured hepatocytes iε also observed.

When the large acidophilic liver reserve cellε are placed in a long-term co-culture with stromal cellε, three populations of cells emerge during the course of culture. The majority of the cultured cells display a typical PC morphology. While a substantial number of the remaining cells are stromal cells, a small yet detectable population of acidophilic reserve cells is invariably present over the course of several weeks. The ability to maintain these hepatic cells long-term, even in a quiescent state, is important for asseεεing the effects of variouε toxicantε on these cells (Michalopouloε and Pitot, 1975, Exp. Cell Reε. 94:70-78; Bissell et al., 1987, J. Clin. Invest. 79:801-812; Guguen-Guilluozo et al., 1983, Exp. Cell Res. 143:47-54). In the co- culture system, cytochrome P-450 enzyme activity is evident in rat hepatocytes for more than 60 dayε of culture. This activity is observed in all diεtinct populations of cells based on FLS vs. SS characteristicε. The smaller cells in these mixed cultureε, which diεplay low to moderate granularity, appear to be endothelial and Kupffer cellε. Both endothelial and Kupffer cellε have been reported to modulate PC cytochrome P-450 activity in vivo and in vitro (Begue et al., 1983, Biochem. Pharmacol. 32:1643-1646; Ratanasavahn et al., 1988, Xenobiotica 18:765). Macrophages have also been reported to exhibit their own detoxifying activities (Wichramasinghe, 1987, Clin. Lab. Hematol. 9:271-280.). In this regard, it was reported recently that the macrophage component of bone marrow stroma was capable of metabolizing EFEE to fluorescein (Naughton et al., 1992, Proc. Soc. Exp. Biol. Med. 201:481-490) . However, the Kupffer cellε in the co-culture εyεtem are conεiderably smaller in size than most of the PC, and it iε primarily the larger cellε that diεplay the greatest cytochrome P-450 activity as manifested by their ability to metabolize fluorescein to EFEE.

The growth of long-term cultured liver PC or liver reserve cells may be enhanced by the addition of various supplements in the culture media. Tissues that normally contain low levels of iron exhibit toxic responseε to iron at higher levelε, but hepatocyteε normally εubεiεt in an iron-rich environment and their intracellular iron stores rapidly diminish in vitro. Thus, the failure of hepatic cells to proliferate in most cultures may be attributed in part to a deficit in iron. Supplementation of the culture medium with iron εaltε and saturated transferrin inhibits albumin production but enhanceε radiothymidine incorporation into hepatocyteε. Mammalian iron εtorage occurs primarily in the liver and spleen. This mineral is necesεary for cell growth, in part because of its role as a co-factor in the activation of ribonucleotide reductase which is needed for DNA synthesiε (Reichard and Ehrenbergen, 1983, Science 221:514-519). In addition, the number of transferrin receptors has been correlated with proliferative activity (Sutherland et al., 1981, Proc. Nat. Acad. Sci. USA 78:4515-4519); these vary during the cell cycle with highest expression occurring in S phase (Yeh et al. , 1982, Exp. Cell Res. 138:429-434). Monoclonal antibody 42/6, which recognizes the transferrin receptor, blocks transferrin binding and inhibits the growth of several cell lines in vitro (Trowbridge et al., 1984, Biochem. Pharmacol. 33:925-932; Mendelsohn et al., 1983, Blood 62:821-826).

Hepatic eythropoietic factor (HEF) , a factor found in the sera of animals with livers regenerating after partial hepatectomy (Naughton et al., 1980, Amer. J. Physiol. 238:E245-E252) , also enhances DNA synthesis of hepatocytes in the culture system described herein. There have been reports on the trophic effects of sera from hepatectomized animals on liver and other tissues of normal animals as well as cultured hepatic cells (Bucher, 1963, Int. Rev. Cytol. 15:245-300). Although the precise mechanisms controlling hepatic regeneration have not been elucidated fully, hepatotropic activity was found in the effluent of regenerating liver that was perfused x situ (Dornfest et al. , 1981, Ann. Clin. Lab. Sci. 11:27-46), suggesting that the liver may contribute to its own regulation. Currently, these factors have not been fully characterized but may all be useful in εupporting long-term growth of liver reεerve cellε and P^C-

5.3. CHARACTERIZATION OF LIVER RESERVE CELLS

Aε εhown by Example 6, infra, liver reεerve cellε can be iεolated and enriched by various procedures. Liver reserve cells have distinct physical characteristics that distinguiεh them from oval cellε or any other hepatocyteε reported thus far. For example, they are larger (> 30 μm in diameter) than typical hepatic PC which are generally in the range of 22-26 μm in diameter, have a low nuclear:cytoplasmic ratio and have 1 or 2 nuclei, each of which has 2-3 prominent nucleoli. When stained with a combination of acidic and basic stains such as Diff-Quick, they primarily retain the acid component of the εtain. In fact, εince all liver PC are highly basophilic, the liver reεerve cellε are stained so faintly that they were originally thought to be "cell ghosts" or dead cells. In order to visualize these cells more clearly, a cytosmear iε usually εtained for longer periodε of time to enhance their uptake of the εtain. Thuε, the term "acidophilic" iε uεed in comparison to other liver cell populations which are highly baεophilic. In addition, unlike oval cellε, they are preεent in normal liver, do not require prior induction with chemical carcinogens, and they either expresε and/or develop into cellε which display the PC-associated functions of albumin secretion and cytochrome P-450 enzyme activity. In culture, they are strongly adherent to plastic and can alεo adhere to other hepatic PC. They have a relatively high mitotic index, capable of dividing every 24 to 28 hourε regularly for ten days to two weeks in monolayer culture, and after thiε period become non-adherent. At thiε stage, the detached cells are morphologically indiεtinguishable from the typical hepatic PC.

Liver reserve cellε may be characterized further by their reactivity with a variety of known cell εurface marker-εpecific monoclonal antibodies. For example, they express low levels of MHC Clasε I antigen, but do not express detectable MHC Clasε II antigen. In addition, liver reserve cells may express other markers which have not yet been identified. Therefore, in order to further characterize these cells, they may be used to generate antibodies against their cell surface molecules. , Alεo within the scope of the invention is the production of polyclonal and monoclonal antibodies which recognize novel antigenic markers expressed by liver reεerve cellε. Such antibodieε may have a variety of uses such as the isolation of liver reεerve cellε by affinity chromatography. Variouε procedures known in the art may be used for the production of antibodieε to liver reserve cells. Various host animals can be immunized by injection with viable liver reserve cells, fixed cells or membrane preparations, including but not limited to rabbits, hamsters, mice, rats, etc. Various adjuvants may be used to increase the immunological responεe, depending on the hoεt species, including but not limited to

Freund's (complete and incomplete), mineral gels such as aluminum hydroxide, εurface active subεtanceε εuch aε lyεolecithin, pluronic polyolε, polyanionε, peotideε, oil emulεionε, keyhole limpet hemocyanin, dinitrophenol, and potentially useful human adjuvants such as BCG (bacille Calmette-Guerin) and Corvnebacterium parvum.

Monoclonal antibodies to novel antigens on liver reεerve cellε may be prepared by uεing any technique which provideε for the production of antibody molecules by continuous cell lines in culture. Theεe include, but are not limited to, the hybridoma technique originally described by Kohler and Milstein (1975, Nature 256. 495-497), and the more recent human B-cell hybridoma technique (Koεbor et al., 1983, Immunology Today 4:72; Cote et al., 1983, Proc. Natl. Acad. Sci. 80:2026-2030) and the EBV-hybridoma technique (Cole et al., 1985, Monoclonal Antibodieε and Cancer Therapy, Alan R. Liεs, Inc., pp. 77-96) . Techniques developed for the production of "chimeric antibodies" by splicing the genes from a mouse antibody molecule of appropriate antigen specificity together with geneε from a human antibody molecule can be uεed (e.σ.. Morriεon et al. , 1984, Proc. Natl. Acad. Sci., 81:6851-6855; Neuberger et al., 1984, Nature, 312:604-608; Takeda et al., 1985, Nature 314:452-454). In addition, techinqueε described for the production of single chain antibodies (U.S. Patent 4,946,778) can be adapted to produce εingle chain antibodies.

Syngeneic, allogeneic, and xenogeneic hostε may be used for injection of liver reserve cellε prepared in viable form, or in fixed form, or as extracted membrane preparations thereof. Monoclonal antibodieε can be εcreened differentially by εelective binding to liver reεerve cellε, but not to other liver PC and εtromal cellε.

Antibody fragmentε which contain the binding site of the molecule may be generated by known techniqueε. For example, εuch fragments include but are not limited to, the F(ab')₂ fragments which can be produced by pepsin digeεtion of the antibody molecule and the Fab fragmentε which can be generated by reducing the diεulfide bridgeε of the F(ab')₂ fragments.

5.4. USES OF LIVER RESERVE CELLS The ability of the large acidophilic cells to proliferate and differentiate into mature biologically active hepatocytes in culture indicates that they are the reserve cells of the liver. As such, they may be particularly useful in transplantation therapy to replace and/or reconεtitute a liver that is genetically deficient, infected by an infectious agent and/or partially destroyed. A major impediment in the current attempts to achieve stable integration of foreign genes in eukaryotic host cells of different organs iε the inability of moεt of theεe cellε to proliferate in vitro. Thiε iε particularly problematic for attempts to insert exogenous genes in liver cells, since hepatocytes do not normally undergo cell division in vitro. Recently, gene transfer studies were performed using hepatocytes isolated from Watanabe heritable hyperlipidemic rabbits, which are widely used as an animal model for familial hypercholesterolemia in humans. Like their human counterparts, the Watanabe rabbit cells contain a genetic deficiency in low density lipoprotein (LDL) receptor, leading to high levels of cholesterol in the circulation and increased incidence of premature coronary artery disease (Wilson et al., 1990, Proc. Natl. Acad. Sci. USA 87:8437). Rabbit hepatocytes were infected with recombinant viruses carrying a functional LDL receptor gene, and shown to cause a temporary amelioration of hyperlipidemia in the genetically deficient rabbitε following tranεplantation. It iε believed that the success rate of this form of therapy can be further augmented if the gene of intereεt can achieve more εtable integration into a population of recipient cellε, which is capable of substantial cell division. Since the liver reεerve cellε proliferate in vitro, eεpecially for longer time periods in the co-culture syεtem deεcribed herein, theεe cellε may be ideal candidates as recipients for the introduction of exogenous genes in culture.

A variety of inborn errorε of metabolism are caused by inherited genetic deficiency in liver cells. These diseases may be treated by transplantation of liver reserve cells carrying functional copies of the correct genes. In brief, this procedure involves isolation of liver reεerve cells from patients afflicted with a particular deficiency, transfer of functional genes into these cells to correct the genetic defect by conventional gene transfer technologies, confirmation of stable integration and expression of the desired gene products, and transplantation of the cells into the patients' own livers for reconstitution. This approach is particularly applicable in situationε where a εingle gene defect iε reεponsible for the diεeaεe and the defective gene has been identified and molecularly cloned; however, it iε not limited only to these conditions. In addition to gene therapy in an autologous setting, liver reserve cells carrying functional geneε may alεo be transplanted into allogeneic HLA-matched individuals. Examples of target geneε and their related liver diεeaεeε that are amenable to thiε form of therapy include, but are not limited to, the LDL receptor gene in familial hypercholeεterolemia, the clotting factor geneε for factorε VIII and IX in hemophilia, the alpha 1-antitrypsin gene in emphysema, the phenylalanine hydroxylase gene in phenylketonuria, the ornithine transcarbamylase gene in hyperammonemia, and complement protein geneε in various forms of complement deficiencies.

The liver is a center of production for many εecretory proteinε. It is anatomically connected with the circulatory system in such a way that allows a efficient release of various proteins into the bloodstream. Therefore, genes encoding proteins that have systemic effects may be inserted into liver reserve cells as opposed to the specific cell types that normally produce them, especially if it is difficult to integrate genes into these cells. For example, a variety of hormone genes or specific antibody geneε may be inεerted into liver reεerve cellε for the secretion of their gene productε into the circulation.

For the practice of the invention, liver reεerve cellε iεolated by the procedureε described in Example 6, infra. are used as recipients in gene transfer experiments. The cells may be grown optimally in a co-culture εyεtem with stromal cellε prior to, during, and after introduction of an exogenouε gene. The proliferative activity of theεe cellε may be enhanced by various iron supple entε deεcribed in Section 5.2., εupra. and in vitro differentiation of theεe cellε may be minimized by the addition of cytokineε in a manner εimilar to the uεe of leukemia inhibitory factor in hematopoietic stem cell cultures. For the introduction of exogenous genes into the cultured reserve cellε, any cloned gene may be transferred using conventional techniques, including, but not limited to, microinjection, transfection and tranεduction. In addition, if the liver reεerve cellε expreεε receptorε for the aεialoglycoprotein, plasmids containing the genes of interest may be conjugated to aεialoglycoprotein and added to cellε to induce uptake and expression (Wu et al., 1991, J. Biol. Chem. 266:14338). Thiε procedure iε more gentle on the recipient cellε.

The preferred method of gene tranεfer utilizes recombinant viruses, such as retroviruses and adenoviruseε. For example, when uεing adenovirus expresεion vectors, a coding sequence may be ligated to an adenovirus transcription/translation control complex, e.g.. the late promoter and tripartite leader sequence. This chimeric gene may then be inserted in the adenovirus genome by in vitro or in vivo recombination. Insertion in a nonesεential region of the viral genome (e.g.. region El or E3) will reεult in a recombinant viruε that iε viable and capable of expreεεing the gene product in infected liver reεerve cellε (e.g.. εee Logan & Shenk, 1984, Proc. Natl. Acad. Sci. USA 81: 3655-3659) . Alternatively, the vaccinia virus 7.5K promoter may be used. (e.g.. see, Mackett et al., 1982, Proc. Natl. Acad. Sci. USA 79: 7415-7419; Mackett et al., 1984, J. Virol. 49: 857-864; Panicali et al., 1982, Proc. Natl. Acad. Sci. USA 79: 4927-4931). Of particular intereεt are vectorε baεed on bovine papilloma viruε which have the ability to replicate aε extrachromoεomal elements (Sarver, et al., 1981, Mol. Cell. Biol. 1: 486). Shortly after entry of thiε DNA into cells, the plasmid replicates to about 100 to 200 copies per cell. Transcription of the inserted cDNA does not require integration of the plasmid into the host's chromosome, thereby yielding a high level of expression. Theεe vectorε can be uεed for εtable expression by including a selectable marker in the plasmid, such as, for example, the neo gene.

Alternatively, the retroviral genome can be modified for use as a vector capable of introducing and directing the expression of any gene of interest in liver reεerve cellε (Cone & Mulligan, 1984, Proc. Natl. Acad. Sci. USA 81:6349-6353). High level expression may also be achieved uεing inducible promoterε, including, but not limited to, the metallothionine IIA promoter and heat εhock promoters. For long-term, high-yield production of recombinant proteins, stable expression is preferred. Rather than using expression vectors which contain viral origins of replication, liver reserve cells can be transformed with a cDNA controlled by appropriate expresεion control elements (e.g.. promoter, enhancer, sequences, transcription terminators, polyadenylation siteε, etc.), and a selectable marker. The selectable marker in the recombinant plasmid confers resiεtance to the εelection and allows cellε to εtably integrate the plaε id into their chromosomes and grow to form foci which in turn can be cloned and expanded into cell lines. For example, following the introduction of foreign DNA, engineered liver cellε may be allowed to grow for 1-2 dayε in an enriched media, and then are εwitched to a selective media. A number of εelection systems may be used, including but not limited to the herpes simplex virus thymidine kinase (Wigler, et al., 1977, Cell 11: 223), hypoxanthine- guanine phosphoriboεyltranεferaεe (Szybalska & Szybalski, 1962, Proc. Natl. Acad. Sci. USA 48: 2026), ^and adenine phosphoribosyltransferase (Lowy, et al., 1980, Cell 22: 817) genes. Also, antimetabolite resistance can be used as the basis of selection for dhfr, which confers resistance to methotrexate (Wigler, et al. , 1980, Proc. Natl. Acad. Sci. USA 77: 3567; O'Hare, et al., 1981, Proc. Natl. Acad. Sci. USA 78: 1527); gpt, which confers reεiεtance to mycophenolic acid (Mulligan & Berg, 1981, Proc. Natl. Acad. Sci. USA 78: 2072; neo, which confers resistance to the aminoglycoside G-418 (Colberre-Garapin, et al., 1981, J. Mol. Biol. 150: 1) ; and hygro, which conferε reεiεtance to hygromycin (Santerre, et al. , 1984, Gene 30: 147) genes. Recently, additional selectable genes have been described, namely trpB, which allows cells to utilize indole in place of tryptophan; hisD, which allows cells to utilize histinol in place of histidine (Hartman & Mulligan, 1988, Proc. Natl. Acad. Sci. USA 85: 8047); and ODC (ornithine decarboxylaεe) which confers resistance to the ornithine decarboxylase inhibitor, 2-(difluoromethyl)-DL-ornithine, DFMO (McConlogue L., 1987, In: Current Communications in Molecular Biology, Cold Spring Harbor Laboratory ed.). The liver reserve cells that have integrated a particular gene aε measured by their expresεion of itε gene product by techniques εuch aε Northern blotε and ELISA, may be tranεplanted into the patientε from whom the cellε are originally derived or into a HLA-matched individual. For HLA-matched allogeneic tranεplantation, the liver reserve cells may not neceεεarily require gene tranεfer prior to tranεplantation. For inεtance, liver reεerve cellε obtained from a donor who possesses a functional gene encoding clotting factor VIII may be used directly by transplantation into a HLA-matched hemophiliac patient. The transplanted cellε will presumably multiply and give rise to mature PC performing normal liver functions, including the production of clotting factor VIII.

In addition to uεing liver reεerve cellε for correcting liver gene defects, theεe cellε may be uεed to replenish the liver parenchyma in the case of hepatic cirhosis, or they may be engineered against liver εpecific infectiouε diseases. For example, uninfected liver reserve cells may be obtained from an early εtage hepatitiε patient and uεed aε recipientε for geneε encoding anti-εenεe RNA that iε complementary to critical replication-related genetic elements of a hepatitis virus. The cells may then be transplanted into the patients to control spread of the virus and restore normal liver function. 6. EXAMPLE: IDENTIFICATION, ISOLATION AND CHARACTERIZATION OF LIVER RESERVE CELLS IN LONG-TERM LIVER CULTURES

6.1. MATERIALS AND METHODS

6.1.1. CELL ISOLATION Cells were produced from male Long-Evans rats at 6-9 weeks of age following sodium pentobarbital anesthesia. An 20 gauge angiocath (Terumo Corp., Japan) was inserted into the hepatic portal vein and secured with two 4-0 ligatures. The liver was perfused with 500 ml of Ca⁺⁺-free buffer 5 (Pertoft and Smedsrod, 1987, In: Cell Separation: Methods and Selected Applications, Vol. 4, T.G. Pretlow II and T.P. Pretlow, editors. Academic Press, New York, pp 1-24) which was delivered at 50 ml/min by a Harvard Instruments (MA) peristalic pump and gently o massaged until uniformly blanched. The liver was placed on a modified Buchner funnel and perfused with a buffer containing Ca⁺⁺ and 0.05 g/dl type IV collagenase (Sigma Chemical Co, MO) in a recirculating system for 15-20 min. The liver waε then tranεferred ₅ to a Petri diεh containing collagenase buffer supplemented with 1.5% bovine serum albumin (BSA) and the hepatocytes were liberated into suspension after the perforation of Gliεεon'ε capsule, filtered through a 185 μm nylon sieve, pelleted by centifrifugation, ₀ and resuεpended in complete medium, DMEM conditioned with 6% FBS and 10% equine εerum and εupplemented with 35 μl glucagon (Sigma #G9261) , 10 μg inεulin (Sigma #14011), 0.25 g glucoεe, and 250 μl hydrocortisone hemisuccinate per 500 ml of medium. 5 Hepatic cells were separated into the following subpopulations using Percoll gradient centrifugation: Parenchymal cells. Cell suspensions were layered atop a solution of 25% Percoll (Pharmacia Inc., NJ) in 10X Dulbecco'ε phosphate buffered εaline (PBS) and spun at 800 x g (10°C) for 10 min. to remove εubcellular debris. Cells were washed, reεuspended in medium, and centrifuged against a 70% Percoll gradient. The pellet of this preparation contains a relatively pure population (>90%) of hepatic PC. The osmolality of this suspension was calculated to be 290 Os using the method of Timonen and Saksela (Timonen and Sakεela, 1980, J. Immunol. Methods 36:285).

Liver reserve cells. A population of large

(≥ 30μm in diameter) , acidophilic cellε which proliferate and differentiate in culture to cellε reεembling mature hepatocyteε waε εeparated aε follows: εingle cell εuεpensions of freshly iεolated liver cellε were centrifuged (500 x g/5 min) and the pellet waε reεuεpended in medium. The cell suspension was layered over 25 ml of a 70% v/v solution of 'neat' Percoll and IX PBS (Sigma Chemical Co., MO) and centrifuged at 800 x g for 10 min. The two lower zones (of 4) were pooled, waεhed, and centrifuged against 25%/50% (v/v/, neat Percoll/lX PBS) discontinuous gradient yielding a distinct interface zone and a pellet. The interface (density = 1.0381 g/ml) consists of about 90% large, lightly acidophilic, mono- or binuclear cells with multiple, prominent nucleoli.

S

Stromal cells. Hepatic fibroblastε, endothelia, and Kupffer cellε were concentrated in the following manner: freεhly isolated liver cells were ₀ centrifuged against a 70% Percoll

(Pharmacia) gradient in 10X PBS (denεity=1.09 g/ml) for 10 min. forming a pellet and a central zone. Cells from the central zone were washed and centrifuged on 5 a 25%/50% Percoll column. The interface zone (density=l.03625 g/ml) contained fibroblastic cells, macrophages, endothelial cells, and occasional peripheral blood leukocytes. 0

6.1.2. NYLON SCREEN CULTURE

15 mm X 60 mm nylon filtration screens

(Tetko, NY) with 210 μm spaceε were treated with 1.0 M acetic acid, waεhed in distilled water, and soaked in ₅ fetal bovine serum (FBS) to enhance cellular attachment. These were placed in Tiεεue Tek εlide chamberε (Nunc, Inc. , IL) and inoculated with 10⁷ liver εtromal cellε which were lifted enzymatically from monolayer culture. Screenε were tranεferred to 25 cm² ₀ flaεkε 18-24 hr later. Within 2 weekε, projectionε of developing εtromal cellε extended acroεε 3 to 4 out of every 5 meεh openingε. Screen cultureε were placed in slide chambers, inoculated with 2-5 x 10⁶ hepatic PC or acidophilic liver reserve cells, and transferred to 25 ₅ cm² flasks after 18-24 hr. Cells were cultured (5% CO₂/35-37°C/>90% humidity) in complete medium. Complete medium replacement was performed 6 times per week. Experimental εupplementε to the medium included saturated transferrin, ferrous εulfate, ferric citrate, and sera from rats with regenerating livers. The latter has been found to contain a factor(ε) that induceε hepatocyte proliferation (Bucher, 1963, Int. Rev. Cytol. 15:245-300; Naughton et al., 1980, Amer. J. Phyεiol. 238:E245-E252) .

6.1.3. ALBUMIN ASSAY Medium waε collected during each feeding and teεted for the preεence of rat albumin uεing the enzyme-linked immunoεorbent assay (ELISA) (Bisεell et al., 1987, J. Clin. Inveεt. 79:801-812). The chromatographically pure rat albumin and peroxidaεe-conjugated εheep anti-rat albumin antibody were purchaεed from Cappel Inc (PA) , and abεorbance at 490 nm waε determined using a kinetic microplate reader (Molecular Devices Inc, CA) . 100 μl of spend medium was added to 96 well plates and stored at 0°C for 12-14 hr. The wells were washed with 0.5% Tween-20 in PBS and non-specific binding sites were blocked with 5.0% BSA in PBS. After washing with 0.5% Tween-20, 100 μl of sheep anti-rat albumin-peroxidase conjugate waε added to each well and incubated for 1 hr at 22°C) . The wells were washed with 0.5% Tween-20 and incubated for 15 min with o-phenylenediamine subεtrate (Cappel Inc., PA). The reaction was stopped and absorbance was measured on an EIA reader. Results were read from a standard curve.

6.1.4. CELL COUNTS Total non-adherent and adherent zone cell counts were determined using an impedance principle cell counter (Counter Electronics, FL) . Differential cell counts were based strictly on the morphology of cells stained using Diff-Quick (Baxter SP, IL) . Cells were scored as parenchymal versus stromal. Phagocytosis of colloidal carbon and reaction with FITC-conjugated antibodies to the vW factor VIII segment were uεed to identify Kupffer cellε and endothelial cellε, reεpectively.

6.1.5. FLOW CYTOMETRY

Phenotypic Analysis. Cellε derived from liver cultureε were reacted on ice with 100 μl mouse monoclonal IgG, polymorphic antibodies to either rat MHC I or MHC II antigens which were conjugated to fluorescein isothiocyanate (FITC) (Serotec Inc, UK) . Control cells were treated with mouse IgGj-FITC alone. The εampleε were analyzed uεing an EPICS C flow cytometer (Coulter Electronics, Hialeah, FL) tuned to a wavelength of 488 nm with the fluorescence gain adjusted to exclude 98% of the control cells. Windows were establiεhed around the variouε cell populationε uεing the forward light εcatter (FLS) vε. εide scatter (SS) two parameter hiεtogram and the percentage of poεitively fluorescent events was determined.

Cytochrome P-450 assay. Freshly isolated hepatocytes, hepatocyteε 24 hr after iεolation, and hepatocyteε derived from εuεpended nylon εcreen cultureε of variouε durations were assayed for cytochrome P-450 monooxygenase activity by flow cytometry. One nM of a 1 μM εtock εolution of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) (Chemical Carcinogen Repoεitory, National Cancer Institute, Kansaε City, MO) in dimethylsulfoxide (DMSO) (Sigma Chem. Co.) was added to cell cultures for 18 hr to induce enzyme activity (Miller, 1983, Anal. Chem. 133:46-57). This non-fluorescent compound was found to be an ideal inducer for this assay. Cells in suspended nylon screen cultureε were lifted uεing a trypεin-collagenaεe mixture (Naughton and Naughton, 1989, Ann. N.Y. Acad. Sci. 554:125-140), pelleted and reεuspended in phosphate buffered saline (PBS) at a density of ~5xl0⁵ cellε/ml, stored on ice for 1 hr, and gradually warmed to 37^βC. Cells were analyzed for evidence to cytochrome P-450 enzyme activity by quantifying incremental fluorescein fluorescence in cellε accumulating ethoxyfluorescein ethyl ester (EFEE) (Miller, 1983, Anal. Chem. 133:46-57; White et al., 1987, Biochem. J. 247:23-28). Cells were incubated with 50 nM EFEE (Molecular Probes, Eugene, OR) in PBS for 5 min at 37°C and examined for green fluoreεcence on a flow cytometer with a 515 nm long-paεε filter and tuned to the 488 nm band. Fluoreεcence waε gated on variouε populations of cells based on differences in FLS vs. SS characteristics and was measured once/minute for up to 15 min in samples maintained at 37^βC. Fluorescein accumulation in cells over time waε indicative of cytochrome P-450 activity (Miller, 1983, Anal. Chem. 13:46-57; White et al., 1987, Biochem. J. 247:23-28).

Cell cycle analysis. Cells disεociated from the adherent zones of suspended nylon screen liver cell cultures were subjected to morphometric analysis to aεcertain the relative percentageε of cellε at different εtages of the cell cycle. Cells were stained for DNA content using a modification of the method of Fried et al (Fried et al., 1978, J. Hiεtochem. Cytochem. 26:921-933) and Taylor (Taylor, 1980, J. Histoche . Cytochem. 28:1021-1024). A staining solution consisting of 0.05 mg/ml of propidium iodide (Sigma Chem. Co. , St. Louis, MO) in distilled water with 1.5% Triton-X-100 at a 1 to 5 (v/v) ratio with complete medium was added and the tubeε were gently agitated uεing an orbital shaker for 2-3 min. RNase (1 mg/ l, 50-75 Kunitz units/ml) (Sigma Chem. Co.) was added to each tube for 3 min. Suspensions were filtered through a 50 μm nylon εcreen. After gating on forward light scatter to eliminate debriε from the histogram, red fluorescence was measured on log scale using the cytofluorograph. Histograms were saved to List Mode and analyzed using the PARA I program (Coulter Electronicε, Hialeah, FL) . The coefficient of variation (CV) of the Gj/G₀ peak of all samples was < 3.0. Statistical Analysis. Flow cytometry measurements were taken in triplicate on εample sizeε of 5,000 (phenotypic analysis) to 10,000 (cell cycle analysiε) events. EFEE to fluorescein conversion was measured as a function of time with 3,000 to 5,000 eventε being sampled per minute. All reεults are expressed at mean + 1 standard error of the mean (sem) . Levelε of εignificance (P) were determined uεing Student'ε T teεt. Data were conεidered εignificant at the 5% level.

6.1.6. IMMUNOFLUORESCENCE

Specimens were fixed in 95% ethanol (24 hr/4°C), dehydrated in absolute ethanol under the εame conditions, and cleared in xylene (8 hr/4°C) prior to embedding in paraffin at 56°C. 7 μm sections were cleared in xylene, dehydrated in a graded series of ethanols, and predigeεted with a bovine teεticular hyaluronidase (Sigma Chem Co., MO) solution at 0.5 mg/ml in 0.1 N sodium acetateacetic acid buffer at pH = 6.0 for 5 min. Slides were blocked prior to the addition of primary antisera with a solution containing 4% goat serum, 0.1% BSA, and 0.1% Tween-20 in 0.1 M NaCl. Polyclonal rabbit primary antibodies to either fibronectin or laminin (Telios 5 Pharmaceuticals, CA) were diluted 1:100 in 0.01 M Tris-Cl at pH = 7.6 containing 1.0% goat εerum and reacted with the εections for 1-3 hr at 22°C. Slides were washed 5 x with 0.1% solution of Tween-20 in 0.1 M NaCl and placed in Tris buffer for 10 min prior to ₀ labelling with goat anti-rabbit IgG conjugated to

R-phycoerythrin (Sigma Chem. Co., MO). Sectionε were counterεtained with propidium iodine for 30 min, mounted with Gelmount, and εtudied on a Nikon Epifluoreεcence microεcope. 5

6.1.7. ELECTRON MICROSCOPY

Adherent zone cellε were dissociated with collagenase, washed, and placed in complete medium for

4-6 hr. Cellε were then pelleted by centrifugation, o fixed for 1 hr in 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.3), washed with buffer, postfixed with oεmium tetroxide, dehydrated in increaεing concentrationε of ethanols and propylene oxide, and embedded in Epon 812. Sections were cut on ₅ an LKB Ultramicrotome (LKB Instruments, MD) , stained with 5% uranyl acetate, and poststained with 0.4% lead citrate. Sections were examined with a JOEL 100C transmission electron microscope.

0

5 6.2. RESULTS The results described in this section are data obtained from experiments utilizing the 5 procedures and techniqueε outlined in Section 6.1., supra. which relate to the iεolation, characterization, culture and function of liver reεerve cellε. When hepatic cellε were iεolated following perfuεion and Percoll denεity gradient ₀ separation, four different morphological categories of cells were observed: 1. medium-large cells which displayed at least two nuclei, prominent nucleoli, and occasional membranous "blebs." These cells adhered to plastic dishes as well as to liver εtromal cells and 5 were localized in the 70% Percoll pellet (FIG. 1); 2. small-medium cells that did not adhere to plastic but which displayed εome ability to attach to εtromal cellε. Cellε of thiε group which did not attach died and underwent autolyεiε within 2-3 dayε of culture. o These cells were found in all separation zones but were most prominent in the overlay medium of the 25%/50% gradient; 3. dead cellε of various sizes which were concentrated in the supernatant of the 70% Percoll εpin, and 4. large (> 30 μm) , faintly εtained ₅ acidophilic cellε with 1-2 nuclei and multiple, prominent nucleoli (FIG. 2) . These appear at the interface of the 25%/50% (neat Percoll/IX PBS) centrifugation, and attach to. either plastic diεhes, nylon screen, or εtromal cells. In addition, other PC ₀ attach to these cells once they are anchored. These large, acidophilic cells proliferate for up to 2 weeks in monolayer culture, detach, and become suεpended in the non-adherent zone, where they are morphologically indistinguishable from freshly pelleted hepatocytes ₅ (FIG. 3) . These large acidophilic cells are believed to be the liver reserve cells. These cells persisted for the entire experimental period when co-cultured with εtroma. The association between PC isolated in the 70% Percoll pellet and hepatic εtromal cellε iε εeen in FIG. 4.

Hepatic εtromal cellε were localized to the lower band at the interface of the 25%:50% diεcontinuous Percoll gradient. Included in this population were Kupffer cells (identified by their ability to phagocytose colloidal carbon and expresε MHC II antigenε) , endothelial cellε (which bound antibodies to vW factor VIII) , fibroblastic cellε, and adipocyte-like cellε. Theεe εtromal cellε form a matrix on the nylon screen template which waε εimilar in εome reεpectε to that seen with liver organ εliceε. Fibronectin and laminin deposition were identified by indirect immunofluoreεcence. The fluoresence intensity of the culture matrix labelled with anti-fibronectin antibody appeared to be greater than that seen with liver organ slices implying an enhanced sytheεiε of this extracellular matrix substance.

The PC which were inoculated onto semi-confluent growth of stromal cells on nylon screens proliferated for 2-3 days in culture and formed clusters of 6-20 cellε in areas which previouεly contained only 1 or 2 cells (FIG. 5A) . Proliferation appeared to cease at this time although the cells remained viable. Hepatic acidophilic cells cultured on suspended nylon screen/stromal templates displayed εimilar early growth patterns (FIG. 5B) but this process appeared to continue until a dense zone of firmly attached parenchymal cell clusters was evident.

Certain media supplements were active on liver cells in culture. In this regard, iron supplements enhanced the percentage of hepatocytes in S phase, as evidenced by propidium iodide cytofluorographic analysis (FIG. 6). Factor(s) found in the hepatic venous sera of rats with regenerating livers after subtotal hepatectomy also enhanced DNA εyntheεiε of hepatocyteε in thiε mixed culture system (FIG. 7).

Acidophilic cells also attached to liver stroma but underwent a higher rate of cluster formation than other PC. ³H-thymidine incorporation by theεe cellε was significantly higher than other populations of hepatocytes in culture and cells in suεpended nylon εcreen co-cultureε continued to incorporate thiε radionuclide for > 2 monthε. Differential counts of enzymatically-removed adherent zone cells indicated an increase in the numbers of parenchymal-like cells with evidence of mitotic figures. Transmiεεion electron microscopic studieε revealed that pelleted adherent zone cellε diεplayed ultrastructural characteristicε εimilar to thoεe seen in freεhly-prepared liver cellε (FIG. 8) . PC which were cultured on εuεpended nylon εcreen/εtromal templates remained viable and synthesized significantly greater quantities of albumin than the PC that were used to initiate the cultures (FIG. 9) . Cells in εuεpended nylon screen co-cultures sustained inducible cytochrome P-450 enzyme activity for up to 60 days as indicated by their ability to transform EFEE to fluorescein (FIG. 10) . This conversion was intracellular since the fluorescence measurements were gated on discrete populations of cells identified by their FLS vs. SS characteristics. Although the three major liver cell populations displayed this activity, highest fluorescence was observed in the larger PC. Arbitrary conversion units were calculated as the product of the percent positive fluoreεcence and peak channel number as described by Miller et al . (Miller, 1983, Anal. Chem. 133:46-57). Peak EFEE to fluoreεcein converεion waε higher in the variouε co-cultureε than in either freεhly isolated liver cellε or one day old liquid cultureε of iεolated hepatocyteε. Although thiε parameter did not appear to be contingent upon the age of the co-culture, differential rates and duration of EFEE conversion were observed in cultures of different ages. Cultured PC manifested a diminished expression (2.6%) of class I MHC antigens when compared to freshly isolated hepatic PC (4.9%). In contrast, no MHC class I antigen expresεion waε detectable on stromal cells and MHC class II epitopes on macrophages were substantially lower than on non-cultured cells (3.5% vs. 9.6%, reεpectively) . FIGS. 11 and 12 demonstrate the relative percentage of three major hepatic cell types in the adherent zones of long-term co-cultureε initiated with highly enriched large acidophilic liver reεerve cell, grown on stromal cell coated nylon screens.

Examples of an in vitro use of the liver cell cultures described above are shown in FIG. 13. The toxic effects of different agents are measured in the co-cultured hepatocyteε.

The preεent invention iε not to be limited in εcope by the exemplified embodiments, which are ' intended as illuεtrationε of individual aεpects of the invention. Indeed, various modifications of the invention in addition to those shown and deεcribed herein will become apparent to those skilled in the art from the foregoing description and accompanying drawings. Such modifications are intended to fall within the scope of the appended claims. All publications cited herein are incorporated by reference in their entirety.

Claims

WHAT IS CLAIMED IS:

1. Isolated liver reserve cellε having a cell εize > 30μm in diameter.

2. The liver reεerve cellε of Claim 1 which are acidophilic when εtained with Diff-Quick.

3. The liver reεerve cellε of Claim 2 which are εtrongly adherent to plaεtic in monolayer culture.

4. The liver reεerve cellε of Claim 3 which are adherent to other εmaller liver parenchymal cells.

5. The liver reserve cells of Claim 4 which have 1 or 2 nuclei, each of which contains 2 or 3 prominent nucleoli.

6. The liver reserve cells of Claim 5 which have mitotic indices εubstantially higher than those of other liver cells at the initiation of in vitro culture aε determined by tritiated thymidine uptake.

7. The liver reεerve cellε of Claim 6 which differentiate to give riεe to cellε capable of albumin production and cytochrome P-450 enzyme activity.

8. A cellular composition comprising a εubstantially homogeneous population of liver reserve cells having a cell size of >.30 μm in diameter.

9. The composition of Claim 8 in which the liver reserve cells are acidophilic when stained with Diff-Quick.

10. The composition of Claim 9 in which the liver reserve cells have mitotic indices substantially higher than those of other liver cellε.

11. The composition of Claim 10 in which the liver reserve cells differentiate to give rise to cells capable of albumin production and cytochrome P- 450 enzyme activity.

12. A method of isolating liver reserve cells compriεing subjecting a mixed population of liver cells to 70% Percoll density gradient centrifugation to obtain thereby four zones, the lower two of which contain the liver reserve cells.

13. The method of Claim 12 in which the cells in the two lower zoneε are pooled and centrifuged against 25%/50% Percoll discontinuous gradient to obtain thereby an interface zone containing cells having a buoyant density of 1.0381 g/ml.

14. The method of Claim 13 further comprising harvesting the cells from the upper portion of the interface zone to obtain a population of at least 80% of liver reserve cellε.

15. A method of uεing liver reεerve cells to reconstitute a mammalian liver comprising transplanting an effective amount of the cells into an individual.

16. A method of using liver reserve cells to correct a genetic deficiency in a mammal comprising tranεplanting an effective amount of the cellε containing an exogenouε gene into an individual.

17. The method of Claim 16 in which the gene encodeε a cell εurface protein.

18. The method of Claim 16 in which the gene encodeε a εecreted protein.